Question About Target Variable (efficiency) Values Exceeding 1 in ABEdeepoff.ipynb

[Rtwick-George-Moses]

We ran the code in cell 8 of ABEdeepoff.ipynb for data cleaning and processing, and noticed that the efficiency column in df_external, which is used as the target variable for the model, contains values greater than 1. Could you please clarify how the formula from the paper(https://www.nature.com/articles/s41467-023-41004-3) was implemented to calculate these values? We're trying to understand if this is expected behavior.

[Muyiyuan]

This is normal. There are on-target and off-target "gRNA - target" designs in our library, if you look at data/ABE_Off_endo.txt, the "Y" value in the off_type column means that the gRNA (seq1 column) is the same as the target (seq2 column). Our model does not predict the true efficiency value. Instead, all off-target efficiencies were normalised to the on-target efficiency (in our paper we call this the off:on-target ratio).
Off:on-target ratio = off-target efficiency / on-target efficiency. For a given gRNA, if the off-target efficiency is greater than the on-target efficiency, the ratio is greater than 1.